Satellite on-orbit autonomous task planning method and system

Abstract

The application discloses a satellite on-orbit autonomous task planning method and system, relates to the technical field of satellite on-orbit autonomous task planning, and significantly improves the anti-interference capability of a satellite in an ionosphere abnormal area through task coding of a directed acyclic graph, a lightweight state snapshot and a dynamic recovery mechanism; a redundant backup path of the DAG structure supports local abnormal jumping of a task sequence and reduces the interruption rate; compressed fingerprint data and majority voting rollback reduce recovery time consumption and avoid response delay caused by full-state reconstruction in a traditional scheme; and dynamic weight adjustment optimizes backup path selection logic and enhances the self-healing capability of the system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of satellite on-orbit autonomous mission planning, and in particular to a satellite on-orbit autonomous mission planning method and system. BACKGROUND

[0002] The low-orbit satellite constellation has important application value in the fields of ocean monitoring and weather observation, and its mission planning system needs to maintain high reliability in a complex space environment. The ionosphere anomaly area near the equator forms a high-energy particle dense area due to the superposition of solar activity and geomagnetic disturbance. When a satellite passes through this area, the on-board computer is easily affected by single event effects, causing abnormal jumps in memory data. The current mainstream satellite autonomous mission planning system adopts a real-time dynamic scheduling strategy, which needs to continuously maintain the integrity of core data structures such as task queues and resource states. However, the continuous interference of the ionosphere anomaly area may cause serious problems such as incorrect instruction sequences and task interruptions.

[0003] Current anti-interference schemes mainly focus on hardware redundancy and software fault tolerance. One type adopts a triple module redundancy (TMR) architecture, which performs three copies of key computing units in parallel and outputs a voting result. Although this can reduce the single-point failure rate, it increases power consumption and computing resource occupation, making it difficult to adapt to the lightweight needs of large-scale constellations. Another type is based on a checkpoint recovery mechanism for dynamic task scheduling, which periodically saves task state snapshots and rolls back to the nearest stable node in the event of an anomaly. However, frequent storage operations exacerbate flash wear and tear, and the recovery process requires the reconstruction of 20-30 task nodes on average, causing delays in responding to critical tasks.

[0004] Recently, some schemes have proposed new algorithms such as online planning and distributed negotiation based on reinforcement learning. Although these algorithms perform well in normal environments, their effectiveness is still not as expected in the ionosphere interference. The irreversible damage to neural network model parameters caused by memory bit flipping leads to the failure of planning strategies. The communication links relied on by multi-satellite collaboration experience a sharp increase in bit error rate during ionosphere disturbance, leading to task allocation conflicts. SUMMARY

[0005] In view of the above existing problems, the present application is proposed.

[0006] The present application provides a satellite on-orbit autonomous mission planning method and system to solve the problem that existing schemes rely on high-energy redundant hardware or inefficient state rollback, making it difficult to deal with multi-bit flipping and communication interference in the ionosphere anomaly area.

[0007] To solve the above technical problems, the present application provides the following technical solutions:

[0008] In a first aspect, the present application embodiment provides a satellite on-orbit autonomous mission planning system, which includes,

[0009] The task scheduling module is used to generate and maintain dynamic task sequences;

[0010] The DAG encoding module, connected to the task scheduling module, encodes the dynamic task sequence into a directed acyclic graph structure containing multiple nodes, where each node corresponds to a single task instruction, and adjacent nodes are provided with a hash check value based on task parameters.

[0011] The status snapshot module interacts with the task scheduling module and compresses the core variables of the current task queue to generate fingerprint data according to a preset period.

[0012] The verification and recovery module is configured to trigger a task path jump or status rollback operation when an abnormal hash verification value or inconsistent fingerprint data is detected.

[0013] The directed acyclic graph structure contains at least one main execution path and two backup paths.

[0014] As a preferred embodiment of the satellite on-orbit autonomous mission planning system described in this invention, in the DAG encoding module, each node contains a mission type identifier, an execution time window, and resource requirement parameters, and the dependency relationship between nodes is established through the parent-child association of the mission type identifier.

[0015] After concatenating the task type identifier, execution time window start value, and resource requirement parameters of adjacent nodes, a unique verification value is generated using a preset hash function.

[0016] As a preferred embodiment of the satellite on-orbit autonomous mission planning system described in this invention, the DAG encoding module defines a directed acyclic graph structure G = (V, E), where the vertex set V = {N1, ..., N}. M}, edge set Each node N i Represented as:

[0017] N i =(t i ,s i ,e i ,r i ),

[0018] Among them, t i s represents the task type identifier of the i-th node. i e represents the starting value of the execution time window for the i-th node. i r represents the termination value of the execution time window for the i-th node. i This represents the resource requirement parameter of the i-th node. The dependency relationship between nodes is established by the parent-child association of task type identifiers: if N i For N j If the parent type is E, then (i,j)∈E;

[0019] For each dependency edge (i, j), concatenate the task type identifier, the start value of the execution time window and the resource requirement parameter of the adjacent nodes in order, apply the SHA256 hash algorithm and take the high 32 bits of the output to generate a check value:

[0020] h i→j = TR 32 (SHA256(t i ||s i ||r i ||t j ||s j ||r j )),

[0021] wherein t i represents the task type identifier of the i-th node, s i represents the start value of the execution time window of the i-th node, r i represents the resource requirement parameter of the i-th node, t j represents the task type identifier of the j-th node, s j represents the start value of the execution time window of the j-th node, r j represents the resource requirement parameter of the j-th node, and h i→j represents the hash check value corresponding to the dependency edge (i, j).

[0022] As a preferred scheme of the satellite on-orbit autonomous task planning system, the backup paths of the directed acyclic graph structure satisfy the following conditions:

[0023] The number of nodes of the main execution path is N, and the first backup path contains at least N-2 nodes overlapping with the main path;

[0024] The starting node of the second backup path is consistent with the third node of the main path, and the end node points to the penultimate node of the main path;

[0025] The hash check values between the nodes of each backup path are independent of the main path.

[0026] As a preferred scheme of the satellite on-orbit autonomous task planning system, the execution logic of the state snapshot module includes:

[0027] Synchronization markers are inserted in the task sequence, and when the task scheduling module processes the synchronization marker, the current satellite power, storage occupancy and task queue pointer are extracted;

[0028] The extracted variables are converted into fixed-length fingerprint data through a preset compression algorithm and stored in an ECC-protected memory partition;

[0029] triggering a majority voting rollback mechanism of the check recovery module when the checksum difference of the fingerprint data is over a threshold value for three times in succession;

[0030] In the process of converting into the fixed-length fingerprint data, the variable extraction rule is:

[0031] The power value is discretized with 0.5% precision and mapped to the integer interval of 0-200;

[0032] The storage occupancy is divided into 10 grades with 10% as a step, and encoded into a 4-bit binary number;

[0033] The task queue pointer and the current orbit period number modulo operation result are spliced into 16-bit data.

[0034] As a preferred scheme of the satellite on-orbit autonomous task planning system, wherein the compression algorithm satisfies:

[0035] The power value is discretized with 0.5% precision and mapped to the integer interval of 0-200;

[0036] The storage occupancy is divided into 10 grades with 10% as a step, and encoded into a 4-bit binary number;

[0037] The task queue pointer and the current orbit period number modulo operation result are spliced into 16-bit data.

[0038] The compression algorithm flow includes:

[0039] The discretized power value is left shifted by 12 bits;

[0040] The storage occupancy is divided into 10 grades with 10% as a step, and encoded into a 4-bit binary number;

[0041] The queue pointer and the orbit period splicing result are superimposed to the low 16 bits.

[0042] As a preferred scheme of the satellite on-orbit autonomous task planning system, wherein the state snapshot module extracts the discretized power value d, the storage occupancy α, the task queue pointer q and the orbit period number T, and generates 24-bit fixed-length fingerprint data F through the following preset compression algorithm:

[0043]

[0044] u=q mod2 5 ,

[0045] v=T mod2 5 ,

[0046] W=(u<<5)∣v,

[0047] F = (b << 16) | (s << 10) | W,

[0048] Wherein, d represents the discretized electric quantity value, a represents the decimal form of the current storage occupancy rate, q represents the task queue pointer, T represents the current orbit period number, b represents the quantized 8-bit electric quantity code, s represents the 6-bit storage occupancy code generated by logarithmic scaling, u represents the low 5 bits of the task queue pointer, v represents the low 5 bits of the orbit period number, W represents the 10-bit intermediate code obtained by left shifting u by 5 bits and performing bitwise OR operation with v, and F represents the final 24-bit fixed-length fingerprint data. represents rounding down, log2 represents logarithm with base 2, mod represents modulo operation, << represents binary left shift, and | represents bitwise OR operation.

[0049] As a preferred scheme of the satellite on-orbit autonomous task planning system, the operation of the check recovery module comprises:

[0050] When the hash check of a single node fails, the subsequent task is continued to be executed by skipping the node along the backup path, and the abnormal node is marked as a to-be-repaired state;

[0051] When the fingerprint data check fails, at least two consistent versions of the last three times of fingerprint data are loaded, and the task queue pointer and the resource state parameter are reconstructed based on the version;

[0052] After the state rollback is completed, the DAG encoding module is requested to rebuild the hash check chain of the affected node.

[0053] As a preferred scheme of the satellite on-orbit autonomous task planning system, the dynamic weight of the backup path is adjusted according to historical bit flip records, and the adjustment is specifically as follows:

[0054] For a node that has a check exception in the past 24 hours, the priority weight of the associated backup path is increased to 1.2 times of the main path;

[0055] When two backup paths associated with the same node both have check exceptions, a third backup path is temporarily generated, and the inter-node hash check value of the path is calculated by using double-bit width.

[0056] In a second aspect, the present application provides a satellite on-orbit autonomous task planning method, comprising:

[0057] Step S1, generating an initial task sequence through a task scheduling module, and converting the initial task sequence into a directed acyclic graph structure with redundant backup;

[0058] Step S2, capturing system states and generating compressed fingerprint data at a preset period during task execution;

[0059] Step S3, when the instruction stream check exception or fingerprint data inconsistency is detected, switch to the backup path to continue execution, and restore the consistency state based on the historical fingerprint copy;

[0060] Step S4, after the satellite leaves the ionosphere abnormal area, reconstruct the hash check chain of the damaged node and optimize the backup path weight distribution.

[0061] The beneficial effects of the present application are: the present application significantly improves the anti-interference ability of the satellite in the ionosphere abnormal area through the task coding of the directed acyclic graph, the lightweight state snapshot and the dynamic recovery mechanism; the redundant backup path of the DAG structure supports the local abnormal jump of the task sequence, reduces the interruption rate; the compressed fingerprint data and the majority voting rollback reduce the recovery time consumption, and avoid the response delay caused by the full state reconstruction in the traditional scheme; the dynamic weight adjustment optimizes the backup path selection logic, and enhances the self-healing ability of the system;

[0062] The present application realizes the instruction stream integrity protection under limited resources, balances the real-time performance and reliability, and meets the long-time autonomous operation requirements of the low-orbit constellation in the extreme space environment. BRIEF DESCRIPTION OF DRAWINGS

[0063] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0064] Figure 1 It is a schematic diagram of the framework of the satellite on-orbit autonomous task planning system in embodiment 1.

[0065] Figure 2 It is a schematic diagram of the flow of the satellite on-orbit autonomous task planning method in embodiment 1. DETAILED DESCRIPTION

[0066] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application will be described in detail in conjunction with the drawings of the specification.

[0067] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited to the specific embodiments disclosed below.

[0068] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0069] Example 1, referring to Figure 1 and Figure 2 This embodiment provides a satellite on-orbit autonomous mission planning system, including:

[0070] The task scheduling module is used to generate and maintain dynamic task sequences;

[0071] The DAG encoding module, connected to the task scheduling module, encodes dynamic task sequences into a directed acyclic graph structure containing multiple nodes, where each node corresponds to a single task instruction, and adjacent nodes are provided with hash check values ​​based on task parameters.

[0072] In the DAG encoding module, each node contains a task type identifier, an execution time window, and resource requirement parameters, and the dependency relationship between nodes is established through the parent-child association of the task type identifier.

[0073] After concatenating the task type identifier, execution time window start value, and resource requirement parameters of adjacent nodes, a unique verification value is generated using a preset hash function (SHA-256 truncated to 32 bits).

[0074] In the DAG encoding module, a directed acyclic graph structure G = (V, E) is defined, where the vertex set V = {N1, ..., N} M}, edge set Each node N i Represented as:

[0075] N i =(t i ,s i ,e i ,r i ),

[0076] Among them, t i s represents the task type identifier of the i-th node. i e represents the starting value of the execution time window for the i-th node. i r represents the termination value of the execution time window for the i-th node. i This represents the resource requirement parameter of the i-th node. The dependency relationship between nodes is established by the parent-child association of task type identifiers: if N i For N j If the parent type is E, then (i,j)∈E;

[0077] For each dependency edge (i, j), the task type identifier, the execution time window start value and the resource requirement parameter of the adjacent nodes are spliced in order, and then the SHA256 hash algorithm is applied and the top 32 bits of the output are truncated to generate a check value:

[0078] h i→j = TR 32 (SHA256(t i ||s i ||r i ||t j ||s j ||r j )),

[0079] wherein t i represents the task type identifier of the i-th node, s i represents the execution time window start value of the i-th node, r i represents the resource requirement parameter of the i-th node, t j represents the task type identifier of the j-th node, s j represents the execution time window start value of the j-th node, r j represents the resource requirement parameter of the j-th node, and h i→j represents the hash check value corresponding to the dependency edge (i, j);

[0080] Specifically, by splicing the key fields of adjacent nodes and truncating the top 32 bits of SHA256, the check strength can be guaranteed, the storage overhead can be reduced, the execution time window start value and the resource requirement parameter are included in the check range, parameter tampering or transmission errors can be captured in real time in chain execution, the dependency relationship is established by the parent-child association of the task type identifier, and the graph traversal and backup path generation are more direct;

[0081] The backup paths of the directed acyclic graph structure satisfy the following conditions:

[0082] The number of nodes of the main execution path is N, and the first backup path contains at least N-2 nodes overlapping with the main path;

[0083] The starting node of the second backup path is consistent with the third node of the main path, and the end node points to the second last node of the main path;

[0084] The hash check values between the nodes of each backup path are independent of the main path;

[0085] The state snapshot module interacts with the task scheduling module, and compresses the core variables of the current task queue to generate fingerprint data at a preset period;

[0086] The execution logic of the state snapshot module includes:

[0087] Inserting synchronization markers in the task sequence, when the task scheduling module processes to the synchronization marker, extracting the current satellite power, storage occupancy and task queue pointer;

[0088] Converting the extracted variables into fixed-length fingerprint data through a preset compression algorithm and storing them in the ECC-protected memory partition;

[0089] When the checksum difference of the fingerprint data exceeds the threshold for three consecutive times, triggering the majority voting rollback mechanism of the check recovery module;

[0090] In the process of converting to fixed-length fingerprint data, the variable extraction rule is:

[0091] The power value is discretized with 0.5% accuracy and mapped to the integer interval of 0-200;

[0092] The storage occupancy is divided into 10 grades with a step of 10%, and encoded as a 4-bit binary number;

[0093] The task queue pointer and the current orbit period number are concatenated to form 16-bit data;

[0094] The compression algorithm satisfies:

[0095] The power value is quantized and encoded using piecewise linear quantization, mapping 0-100% power to an 8-bit binary number;

[0096] The storage occupancy is encoded using logarithmic scaling to generate a 6-bit feature value;

[0097] The task queue pointer and the timestamp are concatenated and subjected to a circular shift operation to generate the remaining bit data, and finally combined into 24-bit fingerprint data;

[0098] The compression algorithm process includes:

[0099] The discretized power value is left shifted by 12 bits;

[0100] The storage occupancy encoding is right shifted by 8 bits and then bitwise ORed with the power value;

[0101] The queue pointer and the orbit period concatenation result are superimposed on the low 16 bits;

[0102] In the state snapshot module, the discretized power value d, storage occupancy α, task queue pointer q and orbit period number T are extracted, and 24-bit fixed-length fingerprint data F is generated through the following preset compression algorithm:

[0103]

[0104] u=q mod2 5 ,

[0105] v=T mod25 ,

[0106] W = (u « 5) | v,

[0107] F = (b « 16) | (s « 10) | W,

[0108] where d represents the discretized electric quantity value, a represents the decimal form of the current storage occupancy, q represents the task queue pointer, T represents the current number of orbit periods, b represents the quantized 8-bit electric quantity code, s represents the 6-bit storage occupancy code generated by logarithmic scaling, u represents the low 5 bits of the task queue pointer, v represents the low 5 bits of the number of orbit periods, W represents a 10-bit intermediate code obtained by left shifting u by 5 bits and performing bitwise OR operation with v, and F represents the final 24-bit fixed-length fingerprint data. represents rounding down, log2 represents the logarithm with base 2, mod represents the modulo operation, « represents binary left shift, and | represents bitwise OR operation.

[0109] The binary operation process includes:

[0110] Calculating the quantization code: The result is 8-bit binary,

[0111] Left shifting 8-bit b by 16 bits to generate the high 8-bit field, which is stored in bit positions 23-16.

[0112] Calculating the logarithmic code: The result is 6-bit binary,

[0113] Left shift alignment: left shift 6-bit s by 10 bits and place it in bit positions 15-10.

[0114] Extracting low bits: u = q mod 2 5 , v = T mod 2 5 , each of which is 5-bit binary.

[0115] Splicing generation: first left shift u by 5 bits to get u « 5, then perform bitwise OR operation with v to get a 10-bit intermediate code W, which is located in bit positions 9-0.

[0116] Merging by bitwise OR operation: F = (b « 16) | (s « 10) | W, forming a 24-bit fixed-length fingerprint.

[0117] Specifically, this step realizes structured and efficient coding by refining binary bit operations, so that each variable occupies a fixed and non-overlapping bit interval in the 24-bit fingerprint. The power value is mapped to 8 bits that can distinguish 256 levels of change, meeting the requirement of capturing small fluctuations. The logarithmic scaling compression of storage occupancy is reduced to 6 bits, which balances large capacity interval and low occupancy precision. The low 5-bit queue pointer and track period are spliced, and the short-time dynamic information of both is completely retained in 10-bit space. Left shift and bitwise OR operation can be efficiently implemented at the hardware or firmware level without floating-point operations, improving embedded execution performance. The 24-bit fixed-length design not only reduces storage and communication burden, but also retains key state features.

[0118] The check recovery module is configured to trigger a task path jump or state rollback operation when a hash check value anomaly or inconsistent fingerprint data is detected.

[0119] The directed acyclic graph structure includes at least one main execution path and two backup paths.

[0120] The operation of the check recovery module includes:

[0121] When a single node hash check fails, skip the node and continue executing subsequent tasks along the backup path, and mark the abnormal node as a to-be-repaired state;

[0122] When the fingerprint data check fails, load at least two consistent versions of the last three fingerprint data, and reconstruct the task queue pointer and resource state parameters based on the version;

[0123] After completing the state rollback, request the DAG encoding module to rebuild the hash check chain of the affected node;

[0124] The dynamic weight of the backup path is adjusted according to the historical bit flip record, specifically:

[0125] For nodes that have experienced check anomalies within the past 24 hours, the priority weight of their associated backup paths is increased to 1.2 times that of the main path;

[0126] When both backup paths associated with the same node experience check anomalies, a third backup path is temporarily generated, and the hash check values between nodes of this path are calculated using double-bit width.

[0127] The embodiment also provides a satellite on-orbit autonomous task planning method, comprising:

[0128] Step S1, generate an initial task sequence through a task scheduling module, and convert it into a directed acyclic graph structure with redundant backups;

[0129] Step S2, during task execution, capture system state and generate compressed fingerprint data at a predetermined period;

[0130] Step S3, when detecting the instruction stream check exception or the fingerprint data inconsistency, switching to the backup path to continue execution, and restoring the consistency state based on the historical fingerprint copy;

[0131] Step S4, after the satellite leaves the ionosphere abnormal area, reconstructing the hash check chain of the damaged node and optimizing the backup path weight distribution.

[0132] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, which should be covered in the scope of the claims of the present application.

Claims

1. A satellite on-orbit autonomous mission planning system, characterized in that, include, The task scheduling module is used to generate and maintain dynamic task sequences; The DAG encoding module, connected to the task scheduling module, encodes the dynamic task sequence into a directed acyclic graph structure containing multiple nodes, where each node corresponds to a single task instruction, and adjacent nodes are provided with a hash check value based on task parameters. The status snapshot module interacts with the task scheduling module and compresses the core variables of the current task queue to generate fingerprint data according to a preset period. The verification and recovery module is configured to trigger a task path jump or status rollback operation when an abnormal hash verification value or inconsistent fingerprint data is detected. The directed acyclic graph structure contains at least one main execution path and two backup paths. The DAG encoding module defines a directed acyclic graph structure. , where vertex set edge set Each node Represented as: ， in, Indicates the first Task type identifier for each node Indicates the first The starting value of the execution time window for each node. Indicates the first The termination value of the execution time window for each node. Indicates the first The resource requirement parameters for each node, and the dependencies between nodes are established by the parent-child association of task type identifiers: if for The parent type, then ; For each dependent edge The task type identifier, execution time window start value, and resource requirement parameters of adjacent nodes are concatenated in sequence, and then the SHA256 hash algorithm is applied to generate a checksum by extracting the high 32 bits of the output. ， in, Indicates the first Task type identifier for each node Indicates the first The starting value of the execution time window for each node. Indicates the first Resource requirements parameters for each node, Indicates the first Task type identifier for each node Indicates the first The starting value of the execution time window for each node. Indicates the first Resource requirements parameters for each node, Represents the corresponding dependent edge The hash check value; In the status snapshot module, the discretized power value is extracted. Storage utilization rate Task queue pointer and orbital period number 24-bit fixed-length fingerprint data is generated using a preset compression algorithm. : ， ， ， ， ， ， in, This represents the discretized electrical quantity value. Represent the current storage occupancy rate in decimal form. Represents the task queue pointer. Indicates the current orbital period number. This represents the quantized 8-bit battery level code. This represents the 6-bit storage occupancy code generated by logarithmic scaling. The lower 5 bits represent the task queue pointer. The lower 5 bits represent the orbital period number. Indicates will After shifting left by 5 bits, and The 10-bit middle code obtained by bitwise OR This represents the final 24-bit fixed-length fingerprint data. Indicates rounding down. Represents the logarithm to the base 2. Modulo operation is represented. Indicates a binary left shift. This indicates a bitwise OR operation.

2. The system as described in claim 1, characterized in that, In the DAG encoding module, each node contains a task type identifier, an execution time window, and resource requirement parameters, and the dependency relationship between nodes is established through the parent-child association of the task type identifier. After concatenating the task type identifier, execution time window start value, and resource requirement parameters of adjacent nodes, a unique verification value is generated using a preset hash function.

3. The system as described in claim 1, characterized in that, The backup path of the directed acyclic graph structure satisfies the following condition: The main execution path has N nodes, and the first backup path contains at least N-2 nodes that overlap with the main path. The starting node of the second backup path is the same as the third node of the main path, and the ending node points to the second-to-last node of the main path. The hash checksum between nodes in each backup path is generated independently of the main path.

4. The system as described in claim 1, characterized in that, The execution logic of the state snapshot module includes: A synchronization marker is inserted into the task sequence. When the task scheduling module processes the synchronization marker, it extracts the current satellite power, storage occupancy rate and task queue pointer. The extracted variables are converted into fixed-length fingerprint data using a preset compression algorithm and stored in an ECC-protected memory partition. When the difference between the checksums of three consecutive fingerprint data exceeds the threshold, the majority vote rollback mechanism of the check recovery module is triggered. The variable extraction rules during the conversion to fixed-length fingerprint data are as follows: The battery level is discretized with a precision of 0.5% and mapped to an integer range of 0-200; Storage utilization is divided into 10 levels in increments of 10%, and the corresponding decimal form of storage utilization is determined based on the level classification. ; The task queue pointer is concatenated with the current orbital cycle digital-to-analog calculation result to form a 16-bit data.

5. The system as described in claim 4, characterized in that, The compression algorithm satisfies: The battery level is encoded using segmented linear quantization, mapping the 0-100% battery level to an 8-bit binary number; The decimal form of the storage occupancy rate obtained from the tiering Logarithmic scaling encoding is used to generate 6-bit feature values; The task queue pointer and the timestamp are concatenated and then subjected to a cyclic shift operation to generate the remaining bits of data, which are finally combined into 24-bit fingerprint data. The compression algorithm process includes: The discretized electrical value is left-shifted by 12 bits. The storage occupancy rate code is right-shifted by 8 bits and then bitwise ORed with the power value. The queue pointer and the orbital period are concatenated and superimposed on the lower 16 bits.

6. The system as described in claim 1, characterized in that, The operation of the verification recovery module includes: When a single node hash verification fails, skip that node along the backup path and continue executing subsequent tasks, and mark the abnormal node as pending repair. When fingerprint data verification fails, load the version that is consistent in at least two of the three most recent fingerprint data, and reconstruct the task queue pointer and resource status parameters based on that version; After completing the state rollback, a request is made to the DAG encoding module to rebuild the hash verification chain of the affected nodes.

7. The system as described in claim 6, characterized in that, The dynamic weight of the backup path is adjusted based on historical bit flip records, specifically as follows: For nodes that have experienced verification anomalies in the past 24 hours, the priority weight of their associated backup paths is increased to 1.2 times that of the primary path; When two backup paths associated with the same node both fail to verify, a third backup path is temporarily generated. The hash verification value between nodes in this path is calculated using double the bit width.

8. A satellite on-orbit autonomous mission planning method, based on the system described in any one of claims 1 to 7, characterized in that, include: Step S1: Generate an initial task sequence through the task scheduling module and convert it into a directed acyclic graph structure with redundant backups; Step S2: During task execution, capture system status and generate compressed fingerprint data according to a preset cycle; Step S3: When an instruction stream verification anomaly or fingerprint data inconsistency is detected, switch to the backup path to continue execution and restore the consistency state based on the historical fingerprint copy. Step S4: After the satellite leaves the ionospheric anomaly zone, reconstruct the hash verification chain of the damaged node and optimize the weight allocation of the backup path.

Images